Abstract:Knowledge graphs are crucial foundations for building intelligent systems, such as question answering and recommendation. However, their performance is hampered by the incompleteness of KGs, so the knowledge graph completion arises to infer whether a triple of the form (head entity, relation, tail entity) is a missing fact. The path-based approach that encodes paths from the head entity to the tail entity for reasoning achieves good performance. Previous work suggests that entity type is beneficial for learning path representations. Nevertheless, the semantics of entities are not captured accurately, as many entities are not typed or loosely typed. In addition, previous methods tend to model paths only from the forward direction but fail to capture new path patterns from the reverse direction (i.e., tail entity to head entity). In this paper, we introduce a structure enhanced path reasoning (SPR) framework to address the above-given problems. First, the model uilizes the structure of entities, i.e., their relational contexts (the relations linked from the given entity), to obtain a reliable path representation that captures correct entity semantics. This information is accessible to all nonisolated entities in all KGs, so that it can compensate the semantics for entities or KGs that have no type available. Second, we leverage the structure of paths to derive their reverse paths, so as to enhance the path representation by additionally encoding the new patterns embedded in them through a dual path encoding method. In order to verify the effectiveness of the proposed methods, we design different architectures based on LSTM and Transformer, respectively. Experimental results on two benchmark datasets, WN18RR, and FB15k-237, show that our approach apparently outperforms state-of-the-art methods on fact prediction task and relation prediction task. Furthermore, extensive experiments illustrate the benefits of enhancing path reasoning by exploiting structure information from entity relational contexts and the dual path encoding method.

Path-based reasoning with K-nearest neighbor and position embedding for knowledge graph completion

Structure Enhanced Path Reasoning for Knowledge Graph Completion

Path-based Reasoning with Constrained Type Attention for Knowledge Graph Completion

Modeling Relation Paths for Knowledge Graph Completion

Reasoning Over Paths via Knowledge Base Completion

Relation Path Embedding in Knowledge Graphs

Representation Learning with Ordered Relation Paths for Knowledge Graph Completion.

Modeling Path Information for Knowledge Graph Completion

Relational Message Passing for Knowledge Graph Completion

Modeling relation paths for knowledge base completion via joint adversarial training

Relational Context Enhanced Dual Path Reasoning for Fact Prediction in Knowledge Graph.

Path Ranking Model for Entity Prediction.

Reasoning Through Memorization: Nearest Neighbor Knowledge Graph Embeddings

Replacing Paths with Connection-Biased Attention for Knowledge Graph Completion

Joint semantics and data-driven path representation for knowledge graph reasoning

Entity-related Paths Modeling for Knowledge Base Completion

ConvHiA: Convolutional Network with Hierarchical Attention for Knowledge Graph Multi-Hop Reasoning

Enhancing Path Information with Reinforcement Learning for Few-shot Knowledge Graph Completion.

Relational Message Passing for Fully Inductive Knowledge Graph Completion

Modeling Relation Path for Knowledge Graph Via Dynamic Projection.

Learning Entity and Relation Embeddings for Knowledge Graph Completion